INDEX

Explanations

statements about the capability of objects or individuals to perform specific actions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "capable" and related phrases indicating ability or potential

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oulos

-0.63

irie

-0.62

 Eisen

-0.61

weed

-0.61

 Crowley

-0.60

abad

-0.59

rez

-0.59

EMS

-0.58

hello

-0.58

 Schwarz

-0.58

POSITIVE LOGITS

of

0.92

 enough

0.86

nesses

0.85

 thereof

0.81

icient

0.77

bodied

0.74

OF

0.71

vana

0.69

idable

0.67

isphere

0.66

Activations Density 0.035%

statements about the capability of objects or individuals to perform specific actions

instances of the word "capable" and related phrases indicating ability or potential

No Comments

No Known Activations